Compensation Averaging Period For Testing

[Guest penman]

I am looking at the 1.401(a)(4)-12 definition of Plan Year Compensation under Section 5 "Special rule for new employees".

Calendar Year Plan

Effective Date = 1/1/04

Comp = Plan Year

Testing Comp = 3 Yr Ave based on Service (rather than Participation)

Testing Date 12/31/04

An participant was hired 7/1/02, 1000 hrs each yr, so there is actually 2.5 years of comp history.

Question: To determine ave comp for testing, can I add the entire comp history and divide by 3 (as long as the method is applied uniformly)?

[SoCalActuary]

How did you determine the benefit? What is your definition of compensation in a year with less than 12 months worked?

In many DB plans, we decide in the document whether we are annualizing or not the first year or last year. I prefer not to annualize, but I only count pay in benefits if the person also gets a year of service credit. In your example, there were three plan years of accrual, so I would average the actual pay during the three year period, (limiting each year to $200k if needed.) Thus 30 months of pay divided by 3 years of testing service would be your testing comp.

However, for testing, you are not tied directly to the compensation definition used in the benefit formula, so my comments are mostly educational. You could use the 30 months / 3 years result for general testing, even if the benefit is not computed the same way.

For DC plan testing, you could use the same approach for each year in which an allocation is made.

[Guest penman]

I typically define ave comp in a DB doc the same way you do. As you said, the doc def puts no restiction on the ave comp testing method. You confirmed my interpretation for ave comp for testing. I realize it's not the only way, but I wanted to confirm it being an acceptable method, thanks.

[AndyH]

Are you testing a db, dc, or a combination? Are you testing on a benefits or contributions basis? This looks like a db but this is in the cross testing board. Please clarify. If it is more than one plan, what is the measurement period?

[Guest penman]

I am testing a combination. Let me answer your question with a question. Per the plan doc, the benefit accrual is based on paricipation. The plan effective date is 1/1/04. Is not my testing service limited to 1 for 2004 whether the measurement period is cur yr or current and all prior yrs? Does the accrual based on participation limit the use of a historical ave comp for determining the EBAR?

What I am trying to do is to take the current year accrual based on current year comp (participation comp) and divide that by the 3 yr ave sal described in the prior post. In that case I suppose the measurement period would be the current yr plus all prior years. Is this a problem?

[SoCalActuary]

You describe a fact pattern that does have only one year of testing service.

If you perform general testing using average comp, I see no problem. It is probably necessary if you are also accruing benefits using compensation before plan entry.

For what it's worth, 415 limits allow use of all years compensation. TH min's will depend on the definition used in your document, and your plan benefit formula can separately and independently choose to use all years compensation. Why not testing also?

[AndyH]

1. It seems to me that if you are testing a DB/DC combo, which you indicated that you are, and the DC plan preceded 2004, then your measurement period does not necessarily have to be the current plan year. I think you could use years benefiting under either plan.

2. If you are trying to average 1 year of accrual over three years of compensation, even if the person only worked 2.5 years, then it is not clear to me whether or not that is permitted. I think not.

1.401(a)(4)-3(e)(2)(i) says that the employees compensation history must be "no shorter than the averaging period". I think that paragraph can be interpreted to prohibit what you want to do. But it is not entirely clear. Absent anything else, I would not do that if you need it to pass.

3. You asked a couple of questions:

Is not my testing service limited to 1 for 2004 whether the measurement period is cur yr or current and all prior yrs?

Not necessarily. See 1 above. Also, clearly no if the DB granted past service.

Does the accrual based on participation limit the use of a historical ave comp for determining the EBAR?

No.

What I am trying to do is to take the current year accrual based on current year comp (participation comp) and divide that by the 3 yr ave sal described in the prior post. In that case I suppose the measurement period would be the current yr plus all prior years

No, I do not agree. The comp averaging period does not define the measurement period. The measurement period may affect your options regarding the comp averaging period, however.

[Guest penman]

Andy,

1. Both plans are effective 1/1/04. I should have stated that.

2. I read that as well and it concerns me. However the a4-12 definition of plan comp under "(5) Special rules for new employees" makes me wonder. Also, I have some seminar notes (2003 EA #405) where Larry D. was describing this same idea except that he was talking about year of termination rather than year of hire. In his example and my example it is still (Full Yr Comp + Full Yr Comp + Partial Year Comp)/3. It is possible I am misreading it though, I don't want to put words in his mouth.

3. I guess I am not clear on "measurement period". Can you please give me a practical example that will help my understanding.

Thanks for your time and help, SoCal also.

[AndyH]

I could be wrong. I am just telling you what I think I am reading, that is all. It appears to me that the special rule for new employees is under Plan Year Compensation. I think in your case Plan Year comp means 2004 comp, NOT average comp. It seems that Plan Year comp is an option if the measurement date is the current plan year or if the plan is an annual accrual plan.

So it looks to me like you can use either 2004 comp or average comp, but if you use average comp the averaging period must start no later than the employment date (or year if you want to push it).

The term year seems to be different because the averaging period must end in the current year, not necessarily on the term date. Does that make sense?

Regarding the measurement period, that is the period of time that you measure the change in the accrued benefit or increase in the account balance. That could be all prior years and the current year. That does not determine testing service, nor does it determine testing comp. Those are separate calculations which may rely to some extent on the measurement period. For example, you cannot use plan year comp unless the measurement period is the current plan year (or the plan is an annual accrual plan).

And testing service is limited to years benefitting or years considered under the formula but limited to the measurement period.

Hope this helps. It is useful to me to dig through this once in a while. I would love to hear confirming or dissenting interpretations.

[Guest penman]

Thanks Andy, very helpful. I hope we get some other opinions as well.